Atkinson's Principles of Clinical Pharmacology

Chapter 1: Introduction to clinical pharmacology

Arthur J. Atkinson, Jr.    Department of Pharmacology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States

Abstract

Clinical pharmacology is defined as the study of drugs in humans. Clinical pharmacologists work both to optimize the use of existing drugs and to develop new medicines. The modern era of clinical pharmacology arose in response to the thalidomide crisis but most adverse drug reactions currently are dose related. Initially, therapeutic drug monitoring used to tailor drug dosage for individual patients. More recently, pharmacogenetics has been used for this purpose and also to aid in drug selection. Progress in this area has been made possible by the development of pharmacokinetics as it relates to the absorption, distribution, and elimination of drugs. Central to this effort has been the assessment of renal function and the elucidation of the pharmacokinetic effects of impaired renal function.

Keywords

Clinical pharmacology; Individualization of drug therapy; Therapeutic drug monitoring; Active drug metabolites; Pharmacokinetics; Glomerular filtration rate estimates

Fortunately a surgeon who uses the wrong side of the

scalpel cuts his own fingers and not the patient; if the

same applied to drugs they would have been investigated

very carefully a long time ago.

Rudolph Buchheim, Beitrage zur Arzneimittellehre, 1849 [1]

Background

Clinical pharmacology can be defined as the study of drugs in humans. Clinical pharmacology often is contrasted with basic pharmacology. Yet applied is a more appropriate antonym for basic [2]. In fact, many basic problems in pharmacology can only be studied in humans. This text will focus on the basic principles of clinical pharmacology. Selected applications will be used to illustrate these principles, but no attempt will be made to provide an exhaustive coverage of applied therapeutics. Other useful supplementary sources of information are listed at the end of this chapter.

Leake [3] has pointed out that pharmacology is a subject of ancient interest but is a relatively new science. Reidenberg [4] subsequently restated Leake’s listing of the fundamental problems with which the science of pharmacology is concerned:

1.The relationship between dose and biological effect.

2.The localization of the site of action of a drug.

3.The mechanism(s) of action of a drug.

4.The absorption, distribution, metabolism, and excretion of a drug.

5.The relationship between chemical structure and biological activity.

These authors agree that pharmacology could not evolve as a scientific discipline until modern chemistry provided the chemically pure pharmaceutical products that are needed to establish a quantitative relationship between drug dosage and biological effect.

Clinical pharmacology has been termed a bridging discipline because it combines elements of classical pharmacology with clinical medicine. The special competencies of individuals trained in clinical pharmacology have equipped them for productive careers in academia, the pharmaceutical industry, and governmental agencies, such as the National Institutes of Health (NIH) and the Food and Drug Administration (FDA). Reidenberg [4] has pointed out that clinical pharmacologists are concerned both with the optimal use of existing medications and with the scientific study of drugs in humans. The latter area includes both evaluation of the safety and efficacy of currently available drugs and development of new and improved pharmacotherapy.

Optimizing use of existing medicines

As the opening quote indicates, the concern of pharmacologists for the safe and effective use of medicine can be traced back at least to Rudolph Buchheim (1820–79), who has been credited with establishing pharmacology as a laboratory-based discipline [1]. In the United States, Harry Gold and Walter Modell began in the 1930s to provide the foundation for the modern discipline of clinical pharmacology [5]. Their accomplishments include the invention of the double-blind design for clinical trials [6], the use of effect kinetics to measure the absolute bioavailability of digoxin and characterize the time course of its chronotropic effects [7], and the founding of Clinical Pharmacology and Therapeutics.

Few drugs have focused as much public attention on the problem of adverse drug reactions as thalidomide, which was first linked in 1961 to catastrophic outbreaks of phocomelia by Lenz in Germany and McBride in Australia [8]. Although thalidomide had not been approved at that time for use in the United States, this tragedy prompted passage in 1962 of the Harris-Kefauver Amendments to the Food, Drug, and Cosmetic Act. This act greatly expanded the scope of the FDA’s mandate to protect the public health. The thalidomide tragedy also provided the major impetus for developing a number of NIH-funded academic centers of excellence that have shaped contemporary clinical pharmacology in this country. These US centers were founded by a generation of vigorous leaders, including Ken Melmon, Jan Koch-Weser, Lou Lasagna, John Oates, Leon Goldberg, Dan Azarnoff, Tom Gaffney, and Leigh Thompson. Colin Dollery and Folke Sjöqvist established similar programs in Europe. In response to the public mandate generated by the thalidomide catastrophe, these leaders quickly reached consensus on a number of theoretically preventable causes that contribute to the high incidence of adverse drug reactions [5]. These include:

1.Inappropriate polypharmacy.

2.Failure of prescribing physicians to establish and adhere to clear therapeutic goals.

3.Failure of medical personnel to attribute new symptoms or changes in laboratory test results to drug therapy.

4.Lack of priority given to the scientific study of adverse drug reaction mechanisms.

5.General ignorance of basic and applied pharmacology and therapeutic principles.

The important observations also were made that, unlike the teratogenic reactions caused by thalidomide, most adverse reactions encountered in clinical practice occurred with drugs that had been in clinical use for a substantial period of time, rather than newly introduced, drugs, and were dose related, rather than idiosyncratic [5, 9, 10].

Recognition of the considerable variation in response of different patients treated with standard drug doses has provided the impetus for the development of what has been called personalized medicine [11] or, more recently, precision medicine [12]. Despite the recent introduction of these terms, they actually describe a continuing story that can be divided into three chapters in which different complementary technologies were developed and are being applied to improve patient therapy by coping with this variability [13]. In the earliest chapter, laboratory methods were developed to measure drug concentrations in patient blood samples and to guide therapy, an approach now termed therapeutic drug monitoring [10]. The routine availability of these measurements then made it possible to apply pharmacokinetic principles in routine patient care to achieve and maintain these drug concentrations within a prespecified therapeutic range. Despite these advances, serious adverse drug reactions (defined as those adverse drug reactions that require or prolong hospitalization, are permanently disabling, or result in death) continue to pose a severe problem and have been estimated to occur in 6.7% of hospitalized patients [14]. Although this figure has been disputed, the incidence of adverse drug reactions probably is still higher than is generally recognized [15]. In the third chapter that is still being written, genetic approaches are being developed and applied to meet both this challenge and to improve the efficacy and safety of drug therapy [16]. Thus pharmacogenetics is being used to identify slow drug-metabolizing patients who might be at increased risk for drug toxicity and rapid metabolizers who might not respond when standard drug doses are prescribed. In a parallel development, pharmacogenomic methods are increasingly used to identify subsets of patients who will either respond satisfactorily or be at increased risk of an adverse reaction to a particular drug.

The fact that most adverse drug reactions occur with commonly used drugs focuses attention on the last of the preventable causes of these reactions: the inadequate training that prescribing physicians receive in pharmacology and therapeutics. Buchheim’s comparison of surgery and medicine is particularly apt in this regard [5]. Most US medical schools provide their students with only a single course in pharmacology that traditionally is part of the second-year curriculum, when students lack the clinical background that is needed to support detailed instruction in therapeutics. In addition, Sjöqvist [17] has observed that most academic pharmacology departments have lost contact with drug development and pharmacotherapy. As a result, students and residents acquire most of their information about drug therapy in a haphazard manner from colleagues, supervisory house staff and attending physicians, pharmaceutical sales representatives, and whatever independent reading they happen to do on the subject. This unstructured process of learning pharmacotherapeutic technique stands in marked contrast to the rigorously supervised training that is an accepted part of surgical training, in which instantaneous feedback is provided whenever a retractor, let alone a scalpel, is held improperly.

Evaluation and development of medicines

Clinical pharmacologists have made noteworthy contributions to the evaluation of existing medicines and development of new drugs. In 1932 Paul Martini published a monograph entitled Methodology of Therapeutic Investigation that summarized his experience in scientific drug evaluation and probably entitles him to be considered the first clinical pharmacologist [18]. Martini described the use of placebos, control groups, stratification, rating scales, and the n of 1 trial design, and emphasized the need to estimate the adequacy of sample size and to establish baseline conditions before beginning a trial. He also introduced the term clinical pharmacology. Gold [6] and other academic clinical pharmacologists also have made important contributions to the design of clinical trials. More recently, Sheiner [19] outlined a number of improvements that continue to be needed in the use of statistical methods for drug evaluation, and asserted that clinicians must regain control over clinical trials in order to ensure that the important questions are being addressed.

Contemporary drug development is a complex process that is conventionally divided into preclinical research and development and a number of clinical development phases, as shown in Fig. 1.1 for small molecule drugs licensed by the US FDA [20]. After a drug candidate is identified and put through in vitro screens and animal testing, an Investigational New Drug application (IND) is submitted to the FDA. When the IND is approved, Phase 1 clinical development begins with a limited number of studies in healthy volunteers or patients. The goal of these studies is to establish a range of tolerated doses and to characterize the drug candidate’s pharmacokinetic properties and initial toxicity profile. If these results warrant further development of the compound, short-term Phase 2 studies are conducted in a selected group of patients to obtain evidence of therapeutic efficacy and to explore patient therapeutic and toxic responses to several dose regimens. These dose-response relationships are used to design longer Phase 3 trials to confirm therapeutic efficacy and document safety in a larger patient population. The material obtained during preclinical and clinical development is then incorporated in a New Drug Application (NDA) that is submitted to the FDA for review. The FDA may request clarification of study results or further studies before the NDA is approved and the drug can be marketed. Adverse drug reaction monitoring and reporting is mandated after NDA approval. Phase 4 studies, conducted after NDA approval, may include studies to support FDA licensing for additional therapeutic indications or over-the-counter (OTC) sales directly to consumers. The development pathway of biologic products is similar and leads to submission of a Biologics License Application (BLA).

Fig. 1.1 The process of new drug development in the United States. (PK indicates pharmacokinetic studies; PD indicates studies of drug effect or pharmacodynamics.) Further explanation is provided in the text. (Modified from Peck CC, Barr WH, Benet LZ, Collins J, Desjardins RE, Furst DE, et al. Opportunities for integration of pharmacokinetics, pharmacodynamics, and toxicokinetics in rational drug development. Clin Pharmacol Ther 1992;51:465–473.)

Although the expertise and resources needed to develop new drugs are primarily concentrated in the pharmaceutical industry, clinical investigators based in academia have played an important catalytic role in championing the development of a number of drugs [21]. For example, dopamine was first synthesized in 1910 but the therapeutic potential of this compound was not recognized until 1963 when Leon Goldberg and his colleagues provided convincing evidence that dopamine mediated vasodilation by binding to a previously undescribed receptor [22]. These investigators subsequently demonstrated the clinical utility of intravenous dopamine infusions in treating patients with hypotension or shock unresponsive to plasma volume expansion. This provided the basis for a small pharmaceutical firm to bring dopamine to market in the early 1970s.

Academically based clinical pharmacologists have a long tradition of interest in drug metabolism. Drug metabolism generally constitutes an important mechanism by which drugs are converted to inactive compounds that usually are more rapidly excreted than the parent drug. However, some drug metabolites have important pharmacologic activity. This was first demonstrated in 1935 when the antibacterial activity of prontosil was found to reside solely in its metabolite, sulfanilamide (Fig. 1.2) [23]. Advances in analytical chemistry over the last 30 years have made it possible to measure on a routine basis plasma concentrations of drug metabolites as well as parent drugs. Further study of these metabolites has demonstrated that several of them have important pharmacologic activity that must be considered for proper clinical interpretation of plasma concentration measurements [24].

Fig. 1.2 Azo-reduction of prontosil to form sulfanilamide and 1,2,4-triaminobenzene.

In some cases, clinical pharmacologists have demonstrated that drug metabolites have pharmacologic properties that make them preferable to marketed drugs. For example, when terfenadine (Seldane), the prototype of nonsedating antihistamine drugs, was reported to cause torsades de pointes and fatality in patients with no previous history of cardiac arrhythmia, Woosley and his colleagues [25] proceeded to investigate the electrophysiologic effects of both terfenadine and its carboxylate metabolite (Fig. 1.3). These investigators found that terfenadine like quinidine, an antiarrhythmic drug with known propensity to cause torsades de pointes in susceptible individuals, blocked the delayed rectifier potassium current. However, terfenadine carboxylate, which actually accounts for most of the observed antihistaminic effects when patients take terfenadine, was found to be devoid of this proarrhythmic property. These findings provided the impetus for commercial development of the carboxylate metabolite as a safer alternative to terfenadine. This metabolite is now marketed as fexofenadine (Allegra).

Fig. 1.3 Chemical structures of terfenadine and its carboxylate metabolite. The acid metabolite is formed by oxidation of the t -butyl side chain of the parent drug.

The potential impact of pharmacogenetics on drug prescribing and development is illustrated by the example of tamoxifen, a selective estrogen receptor modifier that has been used as therapy and recurrence prevention in patients with breast cancer. As shown in Fig. 1.4, tamoxifen is converted by cytochrome P450 (CYP) enzymes to several metabolites that have more potent antiestrogenic activity than the parent compound. Although 4-hydroxy-tamoxifen had been thought to be the primary pharmacologically active tamoxifen metabolite, Flockhart and colleagues [26] demonstrated that endoxifen plasma concentrations averaged more than 10 times those of 4-hydroxy-tamoxifen in women treated with tamoxifen and that both compounds had equal in vitro potency in suppressing breast cancer cell proliferation. Unfortunately, there is still no agreement on the clinical utility of using genotype biomarkers to assess the extent of endoxifen formation to guide breast cancer therapy with tamoxifen, so therapeutic monitoring of endoxifen plasma concentrations is being evaluated for this purpose [27, 28]. These findings also have provided the rationale for current efforts to develop endoxifen as a replacement for tamoxifen that would not be subject to pharmacogenetic variation or drug interactions affecting CYP2D6 activity [29, 30].

Fig. 1.4 Partial metabolic pathway of tamoxifen showing metabolite structures and the CYP enzymes involved. The relative contribution of each metabolic step is indicated by the thickness of the arrows .

Pharmacokinetics

Pharmacokinetics is defined as the quantitative analysis of the processes of drug absorption, distribution, and elimination that determine the time course of drug action in response to an administered drug dose. Pharmacodynamics deals with the mechanism of drug action. Hence, pharmacokinetics and pharmacodynamics constitute two major subdivisions of pharmacology.

Since as many as 70%–80% of adverse drug reactions are dose related [9], our success in preventing these reactions is contingent on our grasp of the principles of pharmacokinetics that provide the scientific basis for dose selection. This becomes critically important when we prescribe drugs that have a narrow therapeutic index. Pharmacokinetics is inescapably mathematical. Although 95% of pharmacokinetic calculations required for clinical application are simple algebra, some understanding of calculus is required to fully grasp the principles of pharmacokinetics.

The concept of clearance

Because pharmacokinetics comprises the first few chapters of this book and figures prominently in subsequent chapters, we will pause here to introduce the clinically most important concept in pharmacokinetics: the concept of clearance. In 1928 Möller et al. [31] observed that, above a urine flow rate of 2 mL/min, the rate of urea excretion by the kidneys is proportional to the amount of urea in a constant volume of blood. They introduced the term clearance to describe this proportionality and defined urea clearance as the volume of blood which one minute’s excretion serves to clear of urea. Since then, creatinine clearance (CLCR) has become most commonly used in clinical practice when renal functional status is directly measured and is calculated from the following equation:

where U is the concentration of creatinine excreted over a certain period of time in a measured volume of urine (V) and P is the serum concentration of creatinine. This is really a first-order differential equation since UV is simply the rate at which creatinine is being excreted in urine (dE/dt). Hence,

If, instead of looking at the rate of creatinine excretion in urine, we consider the rate of change of creatinine in the body (dX/dt), we can write the following equation:

si3_e    (1.1)

Here I is the rate of synthesis of creatinine in the body and CLCR·P is the rate of creatinine elimination. At steady state, these rates are equal and there is no change in the total body content of creatinine (dX/dt = 0), so:

si4_e    (1.2)

This equation explains why it is hazardous to estimate the status of renal function solely from serum creatinine results in patients who have a reduced muscle mass and a concomitant decline in creatinine synthesis rate. For example, creatinine synthesis rate may be substantially reduced in elderly patients, so it is not unusual for serum creatinine concentrations in these patients to remain within normal limits, even though renal function is markedly impaired.

Clinical estimation of renal function

In routine clinical practice, it is not practical to collect the urine samples that are needed to measure creatinine clearance directly. However, creatinine clearance in adult patients can be estimated either from a standard nomogram or from equations such as that proposed by Cockcroft and Gault [32]. For men, creatinine clearance can be estimated from this equation as follows:

   (1.3)

For women, this estimate should be reduced by 15%. By comparing Eq. (1.2) with Eq. (1.3), we see that the terms (140 − age)(weight in kg)/72 simply provide an estimate of the creatinine formation rate in an individual patient.

Since the Cockcroft-Gault equation was introduced, there has been substantial improvement in reducing the variability and analytical bias in automated methods for measuring creatinine concentrations and these measurements are now calibrated to values obtained by isotope dilution mass spectrometry [33]. In addition, the Cockcroft-Gault equation overestimates true glomerular filtration rate (GFR) as measured by inulin clearance because creatinine is secreted by the renal tubule in addition to being filtered at the glomerulus [34]. For these reasons, data from the Modification of Diet in Renal Disease (MDRD) Study have been used by Levey and colleagues [35] to develop a series of equations that more accurately estimate GFR from standardized serum creatinine measurements and other patient characteristics. This group of investigators [36] has used measured renal clearance of iothalamate as a reference to compare GFR estimates and drug dosing recommendations based on the Cockcroft-Gault equation with those obtained using the following 4-variable version of the MDRD Study equation:

Standardized serum creatinine (SCR) measurements were used in both equations without correcting the Cockcroft-Gault equation for this change in analytical precision. Nonetheless, the concordance rates of dosing recommendations for a panel of 15 medications were 88% for the MDRD Study equation and 85% for the Cockcroft-Gault equation when compared with measured GFR. Consequently, the authors recommended basing drug dosing adjustments in patients with impaired renal function on more recent GFR estimating equations rather than on the Cockcroft-Gault equation. Subsequent estimating equations have been developed to extend the prediction range from patients with chronic kidney disease and GFR less than 60 mL/min/1.73 m² to individuals with higher GFR (CKD-EPI) [37] and to incorporate serum concentration of cystatin C, another endogenous GFR marker [38].

Neither the Cockcroft-Gault equation nor the previously described GFR estimating equations can be used to estimate creatinine clearance in pediatric patients because muscle mass has not reached the adult proportion of body weight. Therefore, Schwartz and colleagues [39, 40] developed the following equation to predict creatinine clearance in these patients:

where L is the body length and k varies by age and sex. For children 1–13 years of age, the value of k had been 0.55 but Schwartz et al. [41] have revised this to 0.413, to reflect the introduction of SCR measurements. The original Schwartz formula also recommended discrete values of k for neonates and children under 1 year of age (0.45), and for females (0.57) and males (0.70) between the ages of 13 and 20. Pottel et al. [42] subsequently proposed the following modification of the Schwartz formula in which k for children between 1 and 14 years of age is expressed as the following age-dependent continuous variable:

In all these equations body length is used as a surrogate for muscle mass in order to estimate creatinine generation rate.

The assessment of renal function in the elderly also has been problematic and this dilemma prompted the Berlin Initiative Study (BIS) to develop two separate equations for patients aged 70 years or older that were modeled on iohexol clearance as a GFR reference [43]. The first equation (BIS1) is creatinine based and includes age and sex as additional variables:

A second equation was developed that included measurement of cystatin C but was not deemed as suitable for routine clinical use because of the high cost of cystatin C analysis. A subsequent attempt also has been made to develop an SCR-based equation for estimating GFR as a continuous function across all age groups [44]. However, these latter methods have not been independently validated nor used to guide drug dosage and a recent review of four different methods found that each had limited accuracy [45].

The 2012 clinical guidelines issued by the Kidney Disease: Improving Global Outcomes (KDIGO) group [46] recommend that clinical laboratories report GFR using the 2009 CKD-EPI equations and that an estimating equation incorporating cystatin C be used if confirmation is needed in patients without markers of kidney damage. Fortunately, CKD-EPI equation results can be automatically calculated by clinical laboratories but are normalized to a body surface area of 1.73 m², requiring further calculation to obtain a result that is more consistent with an individual patient’s muscle mass. Given this complexity, the simpler Cockcroft and Gault equation still finds widespread use among clinicians involved in patient care. Unfortunately, estimating equations based on serum creatinine measurements are inaccurate in patients whose renal function is changing rapidly, for example in acute renal failure, [47], and frequently underestimate GFR in trauma and burn patients or in those requiring intensive care in whom augmented renal clearance is common, so measured creatinine clearance must be relied on [48]. On the other hand, creatinine clearance is likely to overestimate renal function in patients with low creatinine production due to cirrhosis, cachexia, or age-related skeletal muscle atrophy [47].

Dose-related toxicity often occurs when impaired renal function is unrecognized

Failure to appreciate that a patient has impaired renal function is a frequent cause of dose-related adverse drug reactions with digoxin and other drugs that normally rely primarily on the kidneys for elimination. As presented in Table 1.1, an audit of patients with high plasma concentrations of digoxin (≥ 3.0 ng/mL) demonstrated that 19 of 44, or 43% of 44 patients with digoxin toxicity had serum creatinine concentrations within the range of normal values, yet had estimated creatinine clearances less than 50 mL/min [49]. Hence, assessment of renal function is essential if digoxin and many other drugs are to be used safely and effectively, and is an important prerequisite for the application of clinical pharmacologic principles to patient care.

Table 1.1

Data from Piergies AA, Worwag EM, Atkinson AJ Jr. A concurrent audit of high digoxin plasma levels. Clin Pharmacol Ther 1994;55:353–358.

Decreases in renal function are particularly likely to be unrecognized in older patients whose creatinine clearance declines as a consequence of aging rather than overt kidney disease. It is for this reason that the Joint Commission on Accreditation of Healthcare Organization has placed the estimation or measurement of creatinine clearance in patients of 65 years of age or older at the top of its list of indicators for monitoring the quality of medication use [50]. Fortunately, computerized laboratory reporting systems have been programmed to automatically report MDRD or CKD-EPI estimates of GFR, a task that was relatively easy to accomplish because these calculations can be performed without access to patient weight. This undoubtedly is an important advance in that it should increase prescriber awareness of a patient’s renal functional status.

Although the developers of the MDRD equation advocated its further use in calculating drug dosage [36], there is a substantial existing body of published dosing guidelines that are based on the Cockcroft-Gault equation. In the final analysis, it may not matter in most cases which equation is used as the basis for adjusting oral doses of many drugs as the accuracy of either equation in estimating renal function generally exceeds the level of adjustment permitted by available oral formulations, or even the accuracy with which tablets can be split.

References

[1] Holmstedt B., Liljestrand G. Readings in pharmacology. Oxford: Pergamon; 1963.

[2] Reidenberg M.M. Attitudes about clinical research. Lancet. 1996;347:1188.

[3] Leake C.D. The scientific status of pharmacology. Science. 1961;134:2069–2079.

[4] Reidenberg M.M. Clinical pharmacology: the scientific basis of therapeutics. Clin Pharmacol Ther. 1999;66:2–8.

[5] Atkinson Jr. A.J., Nordstrom K. The challenge of in-hospital medication use: an opportunity for clinical pharmacology. Clin Pharmacol Ther. 1996;60:363–367.

[6] Gold H., Kwit N.T., Otto H. The xanthines (theobromine and aminophylline) in the treatment of cardiac pain. JAMA. 1937;108:2173–2179.

[7] Gold H., Catell M.K., Greiner T., Hanlon L.W., Kwit N.T., Modell W., et al. Clinical pharmacology of digoxin. J Pharmacol Exp Ther. 1953;109:45–57.

[8] Taussig H.B. A study of the German outbreak of phocomelia: the thalidomide syndrome. JAMA. 1962;180:1106–1114.

[9] Melmon K.L. Preventable drug reactions – causes and cures. N Engl J Med. 1971;284:1361–1368.

[10] Koch-Weser J. Serum drug concentrations as therapeutic guides. N Engl J Med. 1972;287:227–231.

[11] Piquette-Miller M., Grant D.M. The art and science of personalized medicine. Clin Pharmacol Ther. 2007;81:311–315.

[12] Collins F.S., Varmus H. A new initiative on precision medicine. N Engl J Med. 2015;372:793–795.

[13] Atkinson Jr. A.J. Individualization of drug therapy: an historical perspective. Transl Clin Pharmacol. 2014;22:52–54.

[14] Lazarou J., Pomeranz B.H., Corey P.N. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA. 1998;279:1200–1205.

[15] Bates D.W. Drugs and adverse drug reactions. How worried should we be?. JAMA. 1998;279:1216–1217.

[16] Zhang G., Nebert D.W. Personalized medicine: genetic risk prediction of drug response. Pharmacol Ther. 2017;175:75–90.

[17] Sjöqvist F. The past, present and future of clinical pharmacology. Eur J Clin Pharmacol. 1999;55:553–557.

[18] Shelley J.H., Baur M.P. Paul Martini: the first clinical pharmacologist?. Lancet. 1999;353:1870–1873.

[19] Sheiner L.B. The intellectual health of clinical drug evaluation. Clin Pharmacol Ther. 1991;50:4–9.

[20] Peck C.C., Barr W.H., Benet L.Z., Collins J., Desjardins R.E., Furst D.E., et al. Opportunities for integration of pharmacokinetics, pharmacodynamics, and toxicokinetics in rational drug development. Clin Pharmacol Ther. 1992;51:465–473.

[21] Flowers C.R., Melmon K.L. Clinical investigators as critical determinants in pharmaceutical innovation. Nat Med. 1997;3:136–143.

[22] Goldberg L.I. Cardiovascular and renal actions of dopamine: potential clinical applications. Pharmacol Rev. 1972;24:1–29.

[23] Tréfouël J., Tréfouël Mme J., Nitti F., Bouvet D. Activité du p-aminophénylsulfamide sur les infections streptococciques expérimentales de la souris et du lapin. C R Soc Biol (Paris). 1935;120:756–758.

[24] Atkinson Jr. A.J., Strong J.M. Effect of active drug metabolites on plasma level-response correlations. J Pharmacokinet Biopharm. 1977;5:95–109.

[25] Woosley R.L., Chen Y., Freiman J.P., Gillis R.A. Mechanism of the cardiotoxic actions of terfenadine. JAMA. 1993;269:1532–1536.

[26] Stearns V., Johnson M.D., Rae J.M., Morocho A., Novielli A., Bhargava P., et al. Active tamoxifen metabolite plasma concentrations after coadministration of tamoxifen and the selective serotonin reuptake inhibitor paroxetine. J Natl Cancer Inst. 2003;95:1758–1764.

[27] Binkhorst L., Mathijssen R.H.J., Jager A., van Gelder T. Individualization of tamoxifen therapy: much more than just CYP2D6 genotyping. Cancer Treat Rep. 2015;41:289–299.

[28] de Vries Schultink A.H.M., Huitema A.D.R., Beijnen J.H. Therapeutic drug monitoring of endoxifen as an alternative for CYP2D6 genotyping in individualizing tamoxifen therapy. Breast. 2018;42:38–40.

[29] Ahmad A., Shahabuddin S., Sheikh S., Kale P., Manjunath K., Rane R.C., et al. Endoxifen a new cornerstone for breast cancer therapy: demonstration of safety, tolerability, and systemic bioavailability in healthy human subjects. Clin Pharmacol Ther. 2010;88:814–817.

[30] Goetz M.P., Suman V.J., Reid J.M., Northfelt D., Mahr M.A., Ralya A.T., et al. First-in-human phase I study of the tamoxifen metabolite Z-endoxifen in women with endocrine-refractory metastatic breast cancer. J Clin Oncol. 2017;35:3391–3430.

[31] Möller E., McIntosh J.F., Van Slyke D.D. Studies of urea excretion. II. Relationship between urine volume and the rate of urea excretion in normal adults. J Clin Invest. 1928;6:427–465.

[32] Cockcroft D.W., Gault M.H. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16:31–41.

[33] Myers G.L., Miller W.G., Coresh J., Fleming J., Greenberg N., Greene T., et al. Recommendations for improving serum creatinine measurement: a report from the National Kidney Disease Education Program. Clin Chem. 2006;52:5–18.

[34] Bauer J.H., Brooks C.S., Burch R.N. Clinical appraisal of creatinine clearance as a measurement of glomerular filtration rate. Am J Kidney Dis. 1982;2:337–346.

[35] Levey A.S., Coresh J., Greene T., Stevens L.A., Zhang L., Hendriksen S., et al. Using standardized serum creatinine values in the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate. Ann Intern Med. 2006;145:247–254.

[36] Stevens L.A., Nolin T.D., Richardson M.M., Feldman H.I., Lewis J.B., Rodby R., et al. Comparison of drug dosing recommendations based on measured GFR and kidney function estimating equations. Am J Kidney Dis. 2009;54:33–42.

[37] Levey A.S., Stevens L.A., Schmid C.H., Zhang Y., Castro III A.F., Feldman H.I., et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612.

[38] Inker L.A., Schmid C.H., Tighiouart H., Eckfeldt J.H., Feldman H.I., Greene T., et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367:20–29.

[39] Schwartz G.J., Feld L.G., Langford D.J. A simple estimate of glomerular filtration rate in full-term infants during the first year of life. J Pediatr. 1984;104:849–854.

[40] Schwartz G.J., Gauthier B. A simple estimate of glomerular filtration rate in adolescent boys. J Pediatr. 1985;106:522–526.

[41] Schwartz G.J., Muñoz A., Schneider M.F., Mak R.H., Kaskel F., Warady B.A., et al. New equations to estimate GFR in children with CKD. J Am Soc Nephrol. 2009;20:629–637.

[42] Pottel H., Mottaghy F.M., Zaman Z., Martens F. On the relationship between glomerular filtration rate and serum creatinine in children. Pediatr Nephrol. 2010;25:927–934.

[43] Schaeffner E.S., Ebert N., Delanaye P., Frei U., Caedeke J., Jakob O., et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann Intern Med. 2012;157:471–481.

[44] Pottel H., Hoste L., Dubourg L., Ebert N., Schaeffner E., Eriksen B.O., et al. An estimated glomerular filtration rate equation for the full age spectrum. Nephrol Dial Transplant. 2016;31:798–806.

[45] da Silva Selistre L., Rech D.L., de Souza V., Iwaz J., Lemoine S., Dubourg L. Diagnostic performance of creatinine-based equations in estimating glomerular filtration rate in adults 65 years and older. JAMA Intern Med. 2019;179:796–804.

[46] KDIGO Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1–150.

[47] Perrone R.D., Madias N.E., Levey A.S. Serum creatinine as an index of renal function: new insights into old concepts. Clin Chem. 1992;38:1933–1953.

[48] Atkinson Jr. A.J. Augmented renal clearance. Transl Clin Pharmacol. 2018;26:176–181.

[49] Piergies A.A., Worwag E.M., Atkinson Jr. A.J. A concurrent audit of high digoxin plasma levels. Clin Pharmacol Ther. 1994;55:353–358.

[50] Nadzam D.M. A systems approach to medication use. In: Cousins D.M., ed. Medication use. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations; 1998:5–17.

Additional sources of information

General

[Bruton et al., 2018] Bruton L.L., Hilal-Dandan R., Knollman B., eds. Goodman & Gilman’s The pharmacological basis of therapeutics. 13th ed. New York: McGraw-Hill; 2018 This is the standard reference textbook of pharmacology. It contains good introductory presentations of the general principles of pharmacokinetics, pharmacodynamics and therapeutics. Appendix II contains a useful tabulation of the pharmacokinetic properties of many commonly-used drugs.

[Carruthers et al., 2000] Carruthers S.G., Hoffman B.B., Melmon K.L., Nierenberg D.W., eds. Melmon and Morrelli’s Clinical pharmacology. 4th ed. New York: McGraw-Hill; 2000 This is the classic textbook of clinical pharmacology with introductory chapters devoted to general principles and subsequent chapters covering different therapeutic areas. A final section is devoted to core topics in clinical pharmacology.

[Waldman and Terzic, 2009] Waldman S.A., Terzic A., eds. Pharmacology and therapeutics: principles to practice. Philadelphia: Saunders-Elsevier; 2009 This is an introductory textbook that is divided into initial chapters that present pharmacologic principles and later chapter that are devoted to therapeutic applications in a wide number of clinical areas.

Pharmacokinetics

[Gibaldi and Perrier, 1982] Gibaldi M., Perrier D. Pharmacokinetics. 2nd ed. New York: Marcel Dekker; 1982 This is a standard reference in pharmacokinetics and is the one most often cited in the methods section of papers that are published in journals covering this area.

[Rowland and Tozer, 2010] Rowland M., Tozer T.N. Clinical pharmacokinetics and pharmacodynamics: concepts and applications. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2010 This is a well-written book that is very popular as an introductory text.

Drug metabolism

[Pratt and Taylor, 1990] Pratt W.B., Taylor P., eds. Principles of drug action: the basis of pharmacology. 3rd ed. New York: Churchill Livingstone; 1990 This book is devoted to basic principles of pharmacology and has good chapters on drug metabolism and pharmacogenetics.

Drug therapy in special populations

[Evans et al., 1992] Evans W.E., Schentag J.J., Jusko W.J., eds. Applied pharmacokinetics: principles of therapeutic drug monitoring. 3rd ed. Vancouver, WA: Applied Therapeutics; 1992 This book contains detailed information that is useful for individualizing dose regimens of a number of commonly-used drugs.

Drug development

[Gallin et al., 2018] Gallin J., Ognibene F., Johnson L.L. Principles and practice of clinical research. 4th ed. San Diego: Academic Press/Elsevier; 2018 This book is based on a long-standing and successful course taught at the Clinical Center of the National Institutes of Health and provides a practical introduction to the conduct of clinical research.

[Spilker, 1996] Spilker B. Guide to clinical trials. Philadelphia: Lippincott-Raven; 1996 This book contains detailed discussions of many practical topics that are relevant to the process of drug development.

[Yacobi et al., 1993] Yacobi A., Skelly J.P., Shah V.P., Benet L.Z., eds. Integration of pharmacokinetics, pharmacodynamics, and toxicokinetics in rational drug development. New York: Plenum; 1993 This book describes how the basic principles of clinical pharmacology currently are being applied in the process of drug development.

Journals

1 British Journal of Clinical Pharmacology

2 Clinical Pharmacology and Therapeutics

3 Journal of Pharmaceutical Sciences

4 Journal of Pharmacokinetics and Biopharmaceutics

Websites

5 American Society for Clinical Pharmacology and Therapeutics (ASCPT). http://www.ascpt.org/.

6 The American Board of Clinical Pharmacology (ABCP). http://www.abcp.net/.

Chapter 2: Clinical pharmacokinetics

Arthur J. Atkinson, Jr.    Department of Pharmacology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States

Abstract

The premise of the target concentration strategy and its support by therapeutic drug monitoring is that drug concentrations in plasma relate more predictably to drug effects than do drug doses. This is because drug doses are subject to individual variation in drug absorption, distribution, and elimination. Pharmacokinetics can be used to characterize these processes. Drug distribution volume and elimination clearance are the basic pharmacokinetic parameters that govern the disposition of an absorbed drug dose. These two parameters also determine the elimination half-life of a drug. This chapter presents both the mathematical basis and clinical application of pharmacokinetics.

Keywords

Pharmacokinetics; Target concentration strategy; Distribution volume; Elimination clearance; Elimination half-life; Michaelis Menten vs. first-order kinetics; Plateau principle

Pharmacokinetics is a valuable adjunct for prescribing and evaluating patient therapy with many drugs, particularly those that have a narrow therapeutic index, the ratio of toxic/therapeutic drug concentrations. In addition, pharmacokinetics plays an important role in the conduct of both basic and applied pharmacological research, and is an essential component of the drug development process. For most clinical and many other applications, pharmacokinetic analyses can be simplified by representing drug distribution within the body by a single compartment in which drug concentrations are uniform [1]. Clinical application of pharmacokinetics usually entails relatively simple calculations, carried out in the context of what has been termed the target concentration strategy. We shall begin by discussing this strategy.

The target concentration strategy

The rationale for measuring concentrations of drugs in plasma, serum, or blood is that concentration-response relationships are often less variable than are dose-response relationships [2]. This is true because individual variation in the processes of drug absorption, distribution, and elimination affects dose-response relationships, but not the relationship between free (nonprotein-bound) drug concentration in plasma water and intensity of effect (Fig. 2.1). The rationale of therapeutic drug monitoring was first elucidated over 90 years ago when Otto Wuth recommended monitoring bromide levels in patients treated with this drug [3]. However, its more widespread clinical application has been possible only because major advances have been made over the past 50 years in developing analytical methods capable of routinely measuring drug concentrations in patient serum, plasma, or blood samples, and because of increased understanding of basic pharmacokinetic principles [4, 5]

Fig. 2.1 Diagram of factors that account for variability in observed effects when standard drug doses are prescribed. Some of this variability can be compensated for by using plasma concentration measurements to guide dose adjustments.

Because most adverse drug reactions are dose related, therapeutic drug monitoring has been advocated as a means of improving therapeutic efficacy and reducing drug toxicity [6]. Drug concentration monitoring is most useful when combined with pharmacokinetic/pharmacogenetic-based dose selection in an integrated management plan as outlined in Fig. 2.2. This approach to drug dosing is termed the target concentration strategy. Pharmacokinetics has been most useful in estimating initial drug doses, particularly for loading doses and for maintenance doses of drugs that are primarily eliminated by renal excretion, and in making subsequent dose adjustments based on plasma concentration measurements. Recent advances in pharmacogenetics and pharmacogenomics are finding increasing clinical utility in guiding drug selection and in providing dose estimates for drugs that are primarily eliminated by certain metabolic pathways.

Fig. 2.2 Target concentration strategy in which pharmacogenetics (PGX), pharmacokinetics (PK), and drug concentration measurements (TDM) are integral parts of a therapeutic approach that extends from initial drug selection and dose estimation to subsequent patient monitoring and dose adjustment. (Reproduced with permission from Atkinson AJ Jr. Individualization of drug therapy: an historical perspective. Transl Clin Pharmacol 2014;22:52–54.)

Monitoring serum concentrations of digoxin as an example

Given the advanced state of modern chemical and immunochemical analytical methods, the greatest current challenge is the establishment of the range of drug concentrations in blood, plasma, or serum that correlate reliably with therapeutic efficacy or toxicity. This challenge is exemplified by the results shown in Fig. 2.3 that are taken from the attempt by Smith and Haber [7] to correlate serum digoxin levels with clinical manifestations of toxicity. It can be seen that no patient in this study with digoxin levels below 1.6 ng/mL was toxic and that all patients with digoxin levels above 3.0 ng/mL had evidence of digoxin intoxication. However, there was a large intermediate range between 1.6 and 3.0 ng/mL in which patients could be either nontoxic or toxic. Accordingly, laboratory reports of digoxin concentration have traditionally been accompanied by the following guidelines:

Fig. 2.3 Superimposed frequency histograms in which serum digoxin concentrations are shown for 131 patients without digoxin toxicity and 48 patients with electrocardiographic evidence of digoxin toxicity. (Reproduced with permission from Smith TW, Haber E. Digoxin intoxication: the relationship of clinical presentation to serum digoxin concentration. J Clin Invest 1970;49:2377–2386.)

Additional clinical information is often necessary to interpret drug concentration measurements that are otherwise equivocal. Thus Smith and Haber found that all toxic patients with serum digoxin levels less than 2.0 ng/mL had coexisting coronary heart disease, a condition known to predispose the myocardium to the toxic effects of this drug. Conversely, 4 of the 10 nontoxic patients with levels above 2.0 ng/mL were being treated with antiarrhythmic drugs that might have suppressed electrocardiographic evidence of digoxin toxicity.

Although traditional digoxin serum level recommendations were based largely on studies in which digoxin toxicity or intermediate inotropic endpoints were measured, more recent studies have focused on correlating digoxin serum levels with the long-term clinical outcome of patients treated with this drug. The Digitalis Investigation Group trial, in which nearly 1000 patients were enrolled, has forced a major revaluation of digoxin dosing guidelines [8]. The investigators concluded that, compared to placebo, digoxin therapy decreased the need for hospitalization and reduced the incidence of death from congestive heart failure, but not overall mortality. Post hoc analysis of this data indicated that all-cause mortality was only lessened in men whose serum digoxin concentrations ranged from 0.5 to 0.9 ng/mL [9]. Higher levels were associated with progressively greater mortality and did not confer other clinical benefit. Based on the pharmacokinetic properties of digoxin, one would expect levels in this range to be obtained with a daily dose of 0.125 mg. However, most patients with serum digoxin levels in this range were presumed to be taking a 0.25-mg daily digoxin dose, a dose that in patients with normal renal function generally provides a steady-state plasma level of 1.4 ng/mL. In addition, the serum digoxin levels were only measured in a subset of the patients at a single time point, whereas outcomes were followed for a duration of 28–58 months [10].

As a result of subsequent observational studies and meta-analyses, a revised therapeutic range of 0.5–0.9 ng/mL has been recommended with a 56% increase in mortality risk being observed with levels ≥ 1.2 ng/mL [11, 12]. The strongest support for using digoxin is to control rapid heart rate in patients with atrial fibrillation whose blood pressure is only marginally adequate [12]. Digoxin is also recommended for patients with congestive heart failure and reduced cardiac ejection fraction as it has been shown to reduce mortality, morbidity, and hospitalization frequency. However, it has been estimated that only 20% of patients hospitalized for congestive heart failure in recent years were receiving digoxin therapy, whereas in the 1990s more than two-thirds of heart failure patients entering clinical trials were being treated with this drug [13]. In part, this decrease reflects the advent of more effective diuretics and other drugs that unload the left ventricle [12]. However, this may also reflect the fact that appropriate monitoring of digoxin plasma levels and knowledge of pharmacokinetics required to use digoxin safely and effectively is regarded as too much of an inconvenience by most clinicians. So some cardiologists have advocated creating a cadre of dedicated medical heart failure specialists who would have the requisite expertise in these areas [13].

General indications for drug concentration monitoring

Unfortunately, controlled studies documenting the clinical benefit of drug concentration monitoring are limited. In addition, one could not justify concentration monitoring for all prescribed drugs even if this technical challenge could be met. Thus drug concentration monitoring is most helpful for drugs that have a narrow therapeutic index and that have no clinically observable effects that can be easily monitored to guide dose adjustment. Generally accepted indications for measuring drug concentrations are as follows:

1.To evaluate concentration-related toxicity

•Unexpectedly slow drug elimination

•Accidental or purposeful overdose

•Surreptitious drug taking

•Dispensing errors

2.To evaluate lack of therapeutic efficacy

•Patient noncompliance with prescribed therapy

•Poor drug absorption

•Unexpectedly rapid drug elimination

3.To ensure that the dose regimen is likely to provide effective prophylaxis.

4.To use pharmacokinetic principles to guide dose adjustment.

Unfortunately, dose-related adverse reactions still occur frequently with digoxin, phenytoin, and many other drugs for which drug concentration measurements are routinely available. The persistence in contemporary practice of these adverse reactions most likely reflects inadequate understanding of basic pharmacokinetic principles. This is illustrated by the following case history [4]:

A 39-year-old man with mitral stenosis was hospitalized for mitral valve replacement. He had a history of chronic renal failure resulting from interstitial nephritis and was maintained on hemodialysis. His mitral valve was replaced with a prosthesis and digoxin therapy was initiated postoperatively in a dose of 0.25 mg/day. Two weeks later, he was noted to be unusually restless in the evening. The following day, he died shortly after he received his morning digoxin dose. Blood was obtained during an unsuccessful resuscitation attempt, and the measured plasma digoxin concentration was 6.9 ng/mL.

Later in this chapter we will demonstrate that the ostensibly surprising delayed onset of this fatal adverse event was pharmacokinetically consistent with this initial therapeutic decision.

Concepts underlying clinical pharmacokinetics

Pharmacokinetics provides a scientific basis for dose selection, and the process of dose regimen design can be used to illustrate with a single-compartment model the basic concepts of apparent distribution volume (Vd), elimination half-life (t1/2), and elimination clearance (CLE). A schematic diagram of this model is shown in Fig. 2.4 along with the two primary pharmacokinetic parameters of distribution volume and elimination clearance that characterize it.

Fig. 2.4 Diagram of a single-compartment model in which the primary kinetic parameters are the apparent distribution volume of the compartment ( V d ) and the elimination clearance ( CL E ).

Initiation of drug therapy (concept of apparent distribution volume)

Sometimes drug treatment is begun with a loading dose to produce a rapid therapeutic response. Thus a patient with atrial fibrillation might be given a 0.375 mg intravenous loading dose of digoxin as initial therapy to control ventricular rate. The expected plasma concentrations of digoxin are shown in Fig. 2.5. Inspection of this figure indicates that the log plasma-concentration-vs.-time curve eventually becomes a straight line. This part of the curve is termed the elimination phase. By extrapolating this elimination-phase line back to time zero, we can estimate the plasma concentration (C0) that would have occurred if the loading dose were instantaneously distributed throughout the body. Measured plasma digoxin concentrations lie above the back-extrapolated line for several hours because distribution equilibrium actually is reached only slowly after a digoxin dose is administered. This part of the plasma-level-vs.-time curve is termed the distribution phase. This phase reflects the underlying multicompartmental nature of digoxin distribution from the intravascular space to peripheral tissues.

Fig. 2.5 Simulation of plasma ( solid line ) and tissue ( heavy dashed line ) digoxin concentrations after intravenous administration of a 0.375-mg loading dose to a 70-kg patient with normal renal function. A value of 0.7 ng/mL for C 0 is estimated by back-extrapolating ( dotted line ) the elimination-phase plasma concentrations. A value of 536 L for V d is calculated by dividing the administered drug dose by this estimate of C 0 , as shown in the text. Tissue concentrations are referenced to the apparent distribution volume of a peripheral compartment that represents tissue distribution. (Simulation based on pharmacokinetic model of Reuning RH, et al. Role of pharmacokinetics in drug dosage adjustment. I. Pharmacologic effect kinetics and apparent volume of distribution of digoxin. J Clin Pharmacol New Drugs 1973;13:127–141.)

As shown in Fig. 2.5, the back-extrapolated estimate of 0.7 ng/mL for C0 can be used to calculate the apparent volume (Vd(extrap)) of a hypothetical single compartment into which digoxin distribution occurs:

si2_e    (2.1)

In this case,

This distribution volume is much larger than anatomically possible. However, this discrepancy occurs because digoxin has a much higher binding affinity for tissues than for plasma, and the apparent distribution volume is the volume of plasma that would be required to provide the observed dilution of the loading dose. Despite this anomaly, the concept of distribution volume is clinically useful because it defines the relationship between plasma concentration and the total amount of drug in the body. Further complexity arises from the fact that Vd(extrap) is only one of three different distribution volume estimates that we will encounter. Because the distribution process is neglected in calculating this volume, it represents an overestimate of the sum of the volumes of the individual compartments involved in drug distribution.

The time course of the myocardial effects of digoxin parallels its concentration profile in peripheral tissues (Fig. 2.5), so there is a delay between the attainment of peak plasma digoxin concentrations and the observation of maximum inotropic and chronotropic effects. The range of therapeutic and toxic digoxin concentrations has been estimated from observations made during the elimination phase, so blood should not be sampled for digoxin assay until distribution equilibrium is nearly complete. In clinical practice, this means waiting for at least 6 h after a digoxin dose has been administered. In an audit of patients with measured digoxin levels of 3.0 ng/mL or more, it was found that nearly one-third of these levels were not associated with toxicity but reflected procedural error, in that blood was sampled less than 6 h after digoxin administration [14].

For other drugs, such as thiopental [15] or lidocaine [16], the locus of pharmacologic action (termed the biophase in classical pharmacology) is in rapid kinetic equilibrium with the intravascular space. The distribution phase of these drugs represents their somewhat slower distribution from intravascular space to pharmacologically inert tissues, such as skeletal muscle. In this way, the pharmacological effects of single doses of these drugs may be rapidly terminated by the process of distribution even though only a small fraction of the dose has been eliminated from the body. Plasma levels of these drugs reflect therapeutic and toxic effects throughout the dosing interval and blood can be obtained for drug assay without waiting for the elimination phase to be reached.

Continuation of drug therapy (concepts of elimination half-life and clearance)

After starting therapy with a loading dose, maintenance of a sustained therapeutic effect usually necessitates administering additional drug doses to replace the amount of drug that has been excreted or metabolized. Fortunately, the elimination of most drugs is a first-order process in that the rate of drug elimination is directly proportional to the drug concentration in plasma.

Elimination half-life

It is convenient to characterize the elimination of drugs with first-order elimination rates by their elimination half-life, the time required for half an administered drug dose to be eliminated. If drug elimination half-life can be estimated for a patient, it is often practical to continue therapy by administering half the loading dose at an interval of one elimination half-life. In this way, drug elimination can be balanced by drug administration and a steady state maintained from the onset of therapy. Because digoxin has an elimination half-life of 1.6 days in patients with normal renal function, it is inconvenient to administer digoxin at this interval. When renal function is normal, it is customary to maintain digoxin therapy by administering daily doses equal to one-third of the required loading dose.

Another consequence of first-order elimination kinetics is that a constant fraction of total body drug stores will be eliminated in a given time interval. Thus if there is no urgency in establishing a therapeutic effect, the loading dose of digoxin can be omitted and 90% of the eventual steady-state drug concentration will be reached after administering daily doses for a period of time equal to 3.3 elimination half-lives. This is referred to as the Plateau Principle. The classical derivation of this principle is provided later in this chapter, but for now brute force will suffice to illustrate this important concept. Suppose that we elect to omit the 0.375 mg digoxin loading dose shown in Fig. 2.5 and simply begin therapy with a 0.125 mg/day maintenance dose. If the patient has normal renal function, we can anticipate that one-third of the total amount of digoxin present in the body will be eliminated each day and that two-thirds will remain when the next daily dose is administered. As shown in Scheme 2.1, the patient will have digoxin body stores of 0.326 mg just after the fifth daily dose (3.3 × 1.6 day half-life = 5.3 days), and this is 87% of the total body stores that would have been provided by a 0.375 mg loading dose.

Scheme 2.1 Brute force demonstration of drug cumulation when maintenance doses are administered repeatedly.

The solid line in Fig. 2.6 shows ideal matching of digoxin loading and maintenance doses. When the digoxin loading dose (called digitalizing dose in clinical practice) is omitted, or when the loading dose and maintenance dose are not matched appropriately, steady-state levels are reached only asymptotically. However, the most important concept that this figure demonstrates is that the eventual steady state level is determined only by the maintenance dose, regardless of the size of the loading dose. Selection of an inappropriately high digitalizing dose only exposes patients to an interval of added risk without achieving a permanent increase in the extent of digitalization. Conversely, when a high digitalizing dose is required to help control ventricular rate in patients with atrial fibrillation or flutter, a higher than usual maintenance dose also will be required.

Fig. 2.6 Expected digoxin plasma concentrations after administering perfectly matched loading and maintenance doses ( solid line ), no initial loading dose ( bottom dashed line ), or a high loading dose that is large in relation to the subsequent maintenance dose ( upper dashed line ).

Elimination clearance

Just as creatinine clearance is used to quantitate the renal excretion of creatinine, the removal of drugs eliminated by first-order kinetics can be defined by an elimination clearance (CLE). In fact, elimination clearance is the primary pharmacokinetic parameter that characterizes the removal of drugs that are eliminated by first-order kinetics. When drug administration is by intravenous infusion, the eventual steady-state concentration of drug in the body (Css) can be calculated from the following equation (note the similarity to Eq. 1.2), where the drug infusion rate is given by I:

si4_e    (2.2)

When intermittent oral or parenteral doses are administered at a dosing interval, τ, the corresponding equation is:

si5_e    (2.3)

where si1_e is the mean concentration during the dosing interval. Under conditions of intermittent administration, there is a continuing periodicity in maximum (peak) and minimum (‘trough") drug levels so that only a quasi-steady state is reached. However, unless particular attention is directed to these peak and trough levels, no distinction generally is made in clinical pharmacokinetics between the true steady state that is reached when an intravenous infusion is administered continuously and the quasi-steady state that results from intermittent administration.

Because there is a directly proportionate relationship between administered drug dose and steady-state plasma level, Eqs. (2.2), (2.3) provide a straightforward guide to dose adjustment for drugs that are eliminated by first-order kinetics. Thus, to double the plasma level, the dose simply should be doubled. Conversely, to halve the plasma level, the dose should be halved. It is for this reason that Eqs. (2.2), (2.3) are the most clinically important pharmacokinetic equations. Note that, as is apparent from Fig. 2.6, these equations also stipulate that the steady-state level is determined only by the maintenance dose and elimination clearance. The loading dose does not appear in the equations and does not influence the eventual steady-state level.

In contrast to elimination clearance, elimination half-life (t1/2) is not a primary pharmacokinetic parameter because it is determined by distribution volume as well as by elimination clearance:

si7_e    (2.4)

The value of Vd in this equation is not Vd(extrap) but represents a second estimate of distribution volume, referred to as Vd(area) or Vd(β) that generally is estimated from measured elimination half-life and clearance. The similarity of these two estimates of distribution volume reflects the extent to which drug distribution is accurately described by a single-compartment model, and obviously varies from drug to drug [17].

Fig. 2.7 illustrates how differences in distribution volume affect elimination half-life and peak and trough plasma concentrations when the same drug dose is given to two patients with the same elimination clearance. If these two hypothetical patients were given the same nightly dose of a sedative-hypnotic drug for insomnia, si1_e would be the same for both. However, the patient with the larger distribution volume might not obtain sufficiently high plasma levels to fall asleep in the evening, and might have a plasma level that was high enough to cause drowsiness in the morning.

Fig. 2.7 Plasma concentrations after repeated administration of the same drug dose to two hypothetical patients whose elimination clearance is the same but whose distribution volumes differ. The patients have the same si1_e but the larger distribution volume results in lower peak and higher trough plasma levels ( solid line ) than when the distribution volume is smaller ( dashed line ).

Drugs not eliminated by first-order kinetics

Unfortunately, the elimination of some drugs does not follow first-order kinetics. For example, the primary pathway of phenytoin elimination entails initial metabolism to form 5-(parahydroxyphenyl)-5-phenylhydantoin (p-HPPH), followed by glucuronide conjugation (Fig. 2.8). The metabolism of this drug is not first order but follows Michaelis-Menten kinetics because the microsomal enzyme system that forms p-HPPH is partially saturated at phenytoin concentrations of 10–20 μg/mL that are therapeutically effective. The result is that phenytoin plasma concentrations rise hyperbolically as dosage is increased (Fig. 2.9).

Fig. 2.8 Metabolism of phenytoin to form p -HPPH and p -HPPH glucuronide. The first step in this enzymatic reaction sequence is rate limiting and follows Michaelis-Menten kinetics, showing progressive saturation as plasma concentrations rise within the range that is required for anticonvulsant therapy to be effective.

Fig. 2.9 The lines show the relationship between dose and steady-state plasma phenytoin concentrations predicted for two patients who became toxic after initial treatment with 300 mg/day. Measured steady-state plasma concentrations are shown by the solid circles and triangles . The shaded area shows the usual range of therapeutically effective phenytoin plasma concentrations. (Reproduced with permission from Atkinson AJ Jr. Individualization of anticonvulsant therapy. Med Clin North Am 1974;58:1037–1049.)

For drugs eliminated by first-order kinetics, the relationship between dosing rate and steady-state plasma concentration is given by rearranging Eq. (2.3) as follows:

si9_e    (2.5)

The corresponding equation for phenytoin is:

si10_e    (2.6)

where Vmax is the maximum rate of drug metabolism and Km is the apparent Michaelis-Menten constant for the enzymatic metabolism of phenytoin.

Although phenytoin plasma concentrations show substantial interindividual variation when standard doses are administered, they average 10 μg/mL when adults are treated with a 300-mg total daily dose, but rise to an average of 20 μg/mL when the